Quasi Maximum Likelihood Estimation of Large, Approximate Dynamic Factors Models, via the EM Algorithm
Abstract
This paper studies Quasi Maximum Likelihood estimation of dynamic factor models for large panels of time series. Specifically, we consider the case in which the autocorrelation of the factors is explicitly accounted for and therefore the factor model has a state-space form. Estimation of the factors and their loadings is implemented by means of the Expectation Maximization algorithm, jointly with the Kalman smoother. We prove that, as both the dimension of the panel n and the sample size T diverge to infinity, the estimated loadings, factors, and common components are min(√n,√T) -consistent and asymptotically normal. Although the model is estimated under the unrealistic constraint of independent idiosyncratic errors, this mis-specification does not a˙ect consistency. Moreover, we give conditions under which the derived asymptotic distribution can still be used for inference even in case of mis-specifications. Our results are confirmed by a MonteCarlo simulation exercise where we compare the performance of our estimators with Principal Components.
https://arxiv.org/abs/1910.03821
Co-author: Matteo Luciani
About Matteo Barigozzi
Matteo is Associate Professor in Statistics at the London School of Economics and Political Science (LSE). Before joining LSE, he was post-doc researcher at ECARES at the Université libre de Bruxelles. He has an MSc degree in Physics from Università degli Studi di Milano, a MSc in Mathematical Modelling from UNESCO International Centre of Theoretical Physics in Trieste, and a PhD in Economics from Sant’Anna School of Advanced Studies in Pisa.
Matteo's research mainly focuses on high-dimensional time series analysis and specifically on large dynamic factor models with extensions to the non-stationary setting, that is in presence of unit roots and cointegration or of change-points. He is interested also in applications to macroeconomic analysis, as monetary policy making, and financial analysis, as volatility forecasting. He is also working on: sequential testing, models for network data and spectral analysis for modelling mixed frequencies data, non-linearities, and spatial dependencies.
For more information www.barigozzi.eu
See for more Econometric Institute seminars or other Econometric Institute events on the EI website:
